Someone knowledable on memory voltages want to chime in? (also what is PMIC and should I be worried)

[Svedski]

Excuse me for the mess that is the title.

I've had my build for about 6 months, runs fully stable with EXPO on, haven't had a crash/BSOD whatsoever.

I recently noticed that sometimes, usually after 5mins in Windows, sometimes not for a long time, that the RAM sticks report funky numbers in HWINFO (temp jumps of 10*C and then almost instantly back down, which I assume is just wrong reporting - they also go grey and upon restarting HWInfo (and programs like OpenRGB)). the RAM sticks are not detected at all. A reboot of Windows makes them visible again, until it happens again. This happens through literally 6 different installs of Windows, the only software being installed being Realtek Ethernet Drivers, Radeon Adrenalin, Realtek Audio drivers and the AMD chipset drivers.

Apart from funky numbers and dissappearing alltogether, the values titled PMIC undervoltage, PMIC overvoltage and PMIC high temperature get activated (usually right before the RAM dissappears (goes gray in HWinfo). The latter with these PMIC warning worries me, because the build I currently have is, as I've said, fully stable, EXPO on and everything, but it has from what I can tell weird CPU coil whine coming from the motherboard (I've tried 2 mobos they have the same noise) and it worries me that parts will get fried because of this.

I've tested RAM for about ~40mins with AIDA64 and some memtest, both in Windows, and all seems well.

So, my actual question - should I be stressing over these values? Do the voltages and values seem fine? What do the PMIC indicators mean, and are they most likely false values?

Also, RAM temp never exceeds 55*C

screenshot of RAM voltages (stressed for 40mins + idle)

http://imgur.com/DqEBZ1O

PC specs:

7800X3D
GSKILL 2x16GB CL30 F5-6000J3038F16G
850W Superflower Leadex VII PSU
RX7900XT
1x 990PRO M.2 SSD
2x SATA generic SSD's
MSI X670E Tomahawk Wifi


Thanks for any help.

 

[tabascosauz]

The DDR5 monitoring data in HWInfo is known to be really buggy all the time with regards to voltages. If you have already tested it to be stable I wouldn't worry about it.

Mine runs slightly cooler than yours at full load but I wouldn't be alarmed about that stuff as long as you are confident as to where the actual voltages and temps are at.

 

[ir_cow]

That temp sensor is the PCB, not the PMIC or RAM ICs.

I also second notion not worry about 50c. Samsung, Micron and SK Hynix are all rated for 85c (or higher in industrial applications). 1.35v is perfectly fine.

 

[ShrimpBrime]

PMIC is the power controller.

Some ram with RGB can act funky, there are known problems with it.

Best bet is to try a set of memory that doesn't have RGB.

Also, use 1 monitoring software at a time. This will avoid conflicts with poll rates and such.

Otherwise, just use it if it passes memtest and such.

GL!

 

[Svedski]

That temp sensor is the PCB, not the PMIC or RAM ICs.

I also second notion not worry about 50c. Samsung, Micron and SK Hynix are all rated for 85c (or higher in industrial applications). 1.35v is perfectly fine.

Oh no, I'm not worried about 55*C, but it seems weird to me that the PMIC sensors would trigger. I've touched the RAM heatsinks and they don't seem overly hot to me (while they are stressed). Everytime the PMIC sensors trigger the module(s) go undetected. I'd figure if some sort of overvoltage was killing my components (the RAM) the system would be unstable. As I said, everything I throw at the system it handles fine.

I might run an overnight RAM stress test just to be sure, what are your recommendations?

PMIC is the power controller.

Some ram with RGB can act funky, there are known problems with it.

Best bet is to try a set of memory that doesn't have RGB.

Also, use 1 monitoring software at a time. This will avoid conflicts with poll rates and such.

Otherwise, just use it if it passes memtest and such.

GL!

This happens on a fresh install of Windows, I've tried over 5 times. Martin, the HWINFO dev, told me it might be another app interfering with SMBUS. But there is no other app and I've disabled the MSI bloat in the BIOS.

 

[tabascosauz]

PMIC is the power controller.

Some ram with RGB can act funky, there are known problems with it.

Best bet is to try a set of memory that doesn't have RGB.

Also, use 1 monitoring software at a time. This will avoid conflicts with poll rates and such.

Otherwise, just use it if it passes memtest and such.

GL!

I'm pretty sure it's been well confirmed at this point that the HWInfo oddities on DDR5 have nothing to do with RGB. Mine are ripjaws.

 

[ir_cow]

Oh no, I'm not worried about 55*C, but it seems weird to me that the PMIC sensors would trigger. I've touched the RAM heatsinks and they don't seem overly hot to me (while they are stressed). Everytime the PMIC sensors trigger the module(s) go undetected.

I'm not sure what you mean by "go undetected". Like the system is crashes?

 

[Svedski]

I'm not sure what you mean by "go undetected". Like the system is crashes?

As soon as the PMIC under voltage / PMIC overvoltage / PMIC high temp. trigger (they seem to trigger more or less all at once, on both sticks of RAM), it then goes gray in HWinfo(all the values for RAM) and upon restarting HWInfo the RAM is just missing. OpenRGB also can't detect the RAM at that point either.

The system however, remains fully functional, no oddities.

Upon restart, everything is back to normal... until it triggers again.

http://imgur.com/Vb7pO8K

 

[ir_cow]

Hmm I haven't ever paid attention to that before, seems like a new addition to HWinfo. Still 1.35v is perfectly fine. I'm often at 1.5v myself.

 

[tabascosauz]

As soon as the PMIC under voltage / PMIC overvoltage / PMIC high temp. trigger (they seem to trigger more or less all at once, on both sticks of RAM), it then goes gray in HWinfo(all the values for RAM) and upon restarting HWInfo the RAM is just missing. OpenRGB also can't detect the RAM at that point either.

The system however, remains fully functional, no oddities.

Upon restart, everything is back to normal... until it triggers again.

http://imgur.com/Vb7pO8K

Greyed out I think is just HWInfo not receiving [valid] updates from said sensor for some time. Happens a fair bit on some proprietary EC board sensors on some boards of mine, I can't say I've seen it much if at all on my DDR5 though.

Like before, if the system is fine, not unstable, and nothing is awry when the bugs happen then don't worry too much about it.

GPU sensors going blank/gray | HWiNFO® Forum

 

[ShrimpBrime]

I'm pretty sure it's been well confirmed at this point that the HWInfo oddities on DDR5 have nothing to do with RGB. Mine are ripjaws.

You must be assuming I wrote HWInfo has issues with RGB, which I never typed. I meant as in a general sense of things.

So as a courtesy, I did a quick google search to add truth to your claim. Seems disproven?? (I did not read these threads in their entirety.)

HWiNFO64 Causes RAM Lighting issues with Patriot Viper RGB 3600mhz

I've had an issue since I built my computer in July where every time I run HWiNFO64, it causes my RGB light settings on my RAM to reset and begins flickering random lights. While it is easy enough to reset manually, I'd like to start running HWiNFO on start-up and minimized but this forces me...

www.hwinfo.com

Memory Modules Issue

Hi everyone. I have a small issue here. I've already sent a new kit back to newegg and just received a brand new kit of 32gb 4x8gb Corsair Vengeance RGB Pro 3200. When I first installed them, hwinfo64 saw them by their full name, now, and after messing with Corsair iCUE, I only have one that...

www.hwinfo.com

Startup scan makes A2 memory slot to bug on RAM RGB

Hey guys, Build a new system (13900KF, Asus Strix z790-H, Kingston Fury Beast 2x32 6000mt) everything went fine until i noticed that when starting HWiNFO, on startup loading sensors, randomly the memory slot A2 specifically, glitch out on RGB RAM stick to stuck. The default RGB is rainbow...

www.hwinfo.com

Also let it be known that (some motherboards) have RGB controllers that may have some ill effects (not saying strictly and directly to HWInfo64)

This happens on a fresh install of Windows, I've tried over 5 times. Martin, the HWINFO dev, told me it might be another app interfering with SMBUS. But there is no other app and I've disabled the MSI bloat in the BIOS.

I'm going to guess one of the memory sticks has a bad PMIC and causing this issue. Really won't know unless you try different memory sticks, not just another windows install. PMIC is located on the memory stick for controlling the voltages.

Yes, some softwares can be conflicting with drivers and sometimes lack thereof.

I could suggest to update the bios, but being keen, I bet you've tried that already. (bios updates sometimes add memory support)

 

[Svedski]

Ram readings issue

so i got this issue with my ram, i get all different reading on HWiNFO64, Aida and CPUZ, also after using the pc for a couple of minutes the readings are gone on all softwares and i get weird things like the screenshot i'm sharing... is this a bug? or maybe the sensors are failling? i stressed...

www.hwinfo.com

It seems I'm not the only one having just this issue. But this seems to have been fixed according to one user over a year ago. My BIOS is fully up to date.

You must be assuming I wrote HWInfo has issues with RGB, which I never typed. I meant as in a general sense of things.

So as a courtesy, I did a quick google search to add truth to your claim. Seems disproven?? (I did not read these threads in their entirety.)

HWiNFO64 Causes RAM Lighting issues with Patriot Viper RGB 3600mhz

I've had an issue since I built my computer in July where every time I run HWiNFO64, it causes my RGB light settings on my RAM to reset and begins flickering random lights. While it is easy enough to reset manually, I'd like to start running HWiNFO on start-up and minimized but this forces me...

www.hwinfo.com

Memory Modules Issue

Hi everyone. I have a small issue here. I've already sent a new kit back to newegg and just received a brand new kit of 32gb 4x8gb Corsair Vengeance RGB Pro 3200. When I first installed them, hwinfo64 saw them by their full name, now, and after messing with Corsair iCUE, I only have one that...

www.hwinfo.com

Startup scan makes A2 memory slot to bug on RAM RGB

Hey guys, Build a new system (13900KF, Asus Strix z790-H, Kingston Fury Beast 2x32 6000mt) everything went fine until i noticed that when starting HWiNFO, on startup loading sensors, randomly the memory slot A2 specifically, glitch out on RGB RAM stick to stuck. The default RGB is rainbow...

www.hwinfo.com

Also let it be known that (some motherboards) have RGB controllers that may have some ill effects (not saying strictly and directly to HWInfo64)


I'm going to guess one of the memory sticks has a bad PMIC and causing this issue. Really won't know unless you try different memory sticks, not just another windows install. PMIC is located on the memory stick for controlling the voltages.

Yes, some softwares can be conflicting with drivers and sometimes lack thereof.

I could suggest to update the bios, but being keen, I bet you've tried that already. (bios updates sometimes add memory support)

Yeah as stated above I have updated (and downgraded... and updated again) the BIOS. No change. Do you think I could try swapping out one stick at a time to see if the values change / sensors bug out? I can't say I have any spare DDR5 RAM to test on this puppy.

 

[tabascosauz]

You must be assuming I wrote HWInfo has issues with RGB, which I never typed. I meant as in a general sense of things.

So as a courtesy, I did a quick google search to add truth to your claim. Seems disproven?? (I did not read these threads in their entirety.)

HWiNFO64 Causes RAM Lighting issues with Patriot Viper RGB 3600mhz

I've had an issue since I built my computer in July where every time I run HWiNFO64, it causes my RGB light settings on my RAM to reset and begins flickering random lights. While it is easy enough to reset manually, I'd like to start running HWiNFO on start-up and minimized but this forces me...

www.hwinfo.com

Memory Modules Issue

Hi everyone. I have a small issue here. I've already sent a new kit back to newegg and just received a brand new kit of 32gb 4x8gb Corsair Vengeance RGB Pro 3200. When I first installed them, hwinfo64 saw them by their full name, now, and after messing with Corsair iCUE, I only have one that...

www.hwinfo.com

Startup scan makes A2 memory slot to bug on RAM RGB

Hey guys, Build a new system (13900KF, Asus Strix z790-H, Kingston Fury Beast 2x32 6000mt) everything went fine until i noticed that when starting HWiNFO, on startup loading sensors, randomly the memory slot A2 specifically, glitch out on RGB RAM stick to stuck. The default RGB is rainbow...

www.hwinfo.com

Also let it be known that (some motherboards) have RGB controllers that may have some ill effects (not saying strictly and directly to HWInfo64)

Talking specifically about the values occasionally spiking on DDR5. 2 of those threads are DDR4, and the last one doesn't address the behaviour here. I wasn't talking about RGB, but believe me I'm no fan of RGB either and no stranger to the problems it can cause.

 

[ShrimpBrime]

Talking specifically about the values occasionally spiking on DDR5. 2 of those threads are DDR4, and the last one doesn't address the behaviour here.

I spent as much time as it takes for Google to query a response. 0.02865312s

And here is Copy paste of the AI overview. Enjoy.

Spoiler: AI overview to the question: Memory RGB causes issues with hwinfo64

Memory RGB software can sometimes conflict with HWiNFO64, potentially causing issues like crashes, incorrect readings, or RGB lighting malfunctions, especially when accessing the same resources (like SMBus).

Here's a breakdown of the common issues and potential solutions:
Common Issues:

• Crashes/Blue Screens:
  HWiNFO64, when running alongside RGB control software (like iCUE, OpenRGB, etc.), can sometimes lead to system instability and crashes.
  
  
• Incorrect Sensor Readings:
  The RGB software might interfere with HWiNFO64's ability to accurately read memory temperatures or other sensor data.
  
  
• RGB Lighting Issues:
  HWiNFO64 can sometimes cause RGB lights to turn off, malfunction, or become unresponsive, especially when accessing the same SMBus devices.
  
  
• SPD Data Overwrite:
  Some RGB control software, if not well-written, might overwrite the Serial Presence Detect (SPD) data on the memory modules, leading to incorrect memory detection or other issues.

Potential Solutions:

• Disable RGB Control Software:
  Temporarily disabling the RGB control software when using HWiNFO64 can help isolate the issue and determine if it's the source of the conflict.
  
  
• Examine HWiNFO64 Settings:
  • SMBus/I2C Tab: In the HWiNFO64 settings, navigate to the "SMBus / I2C" tab and try excluding specific SMBus devices that might be causing conflicts.
    
    
  • Safety Settings: In the "Safety" tab, check if there are any checkboxes related to specific RGB software (like Corsair Link) that might be causing conflicts and uncheck them.
    
    
  • Disable DIMM Sensors: As a last resort, try disabling the monitoring of DIMM sensors in HWiNFO64 to see if that resolves the issue.
• Update RGB Software:
  Ensure you're using the latest version of your RGB control software, as older versions might have compatibility issues with HWiNFO64.
  
  
• Check for Conflicts:
  If you're using multiple RGB control software, try disabling them one by one to identify which one is causing the conflict.
  
  
• Disable SMBus Support:
  As a troubleshooting step, you can try disabling SMBus support in HWiNFO64 to see if that resolves the issue.
  
  
• Consider Fixed RGB Patterns:
  If possible, try running your RAM RGB with a fixed pattern (without the control software) to see if that avoids the conflict.
  
  
• Start HWiNFO Multiple Times:
  If you're experiencing issues with memory module detection, try starting HWiNFO64 multiple times to see if the problem persists.

Important Notes:

• SMBus Conflicts:
  Many RGB controllers use the SMBus protocol to communicate with the memory modules, and HWiNFO64 also uses this protocol for monitoring. This can lead to conflicts if both applications try to access the same devices at the same time.
  
  
• Not All RGB Software is Equal:
  Some RGB control software is better written and more compatible with other tools than others. If you're experiencing problems, consider switching to a different RGB software.
  
  
• Hardware Issues:
  In rare cases, the issue might be with the hardware itself (e.g., a faulty memory module or motherboard), but in most cases, it's a software conflict.

 

[tabascosauz]

I spent as much time as it takes for Google to query a response. 0.02865312s

And here is Copy paste of the AI overview. Enjoy.

Why are you copy pasting something from Gemini like it adds something to the discussion? Is it still April 1st? lol

I wasn't intending to start a fight or accuse you of anything. Just saying that the VDD values spiking in HWInfo (and thus triggering overvoltage flag) are not anything new for DDR5. Which is what OP initially was saying.

 

[freeagent]

I think its kinda neat that people feed DDR5 the same voltage as DDR4. I guess they run the same voltage.. so it makes sense..

Sorry.. just connecting dots..

 

[tabascosauz]

@Svedski just crossed my mind, have you been keeping chipset drivers up to date while you've been updating BIOS/AGESA? There seem to be some changes afoot and it's not every day I see MSI put in BIOS notes that a certain BIOS revision should be paired with a specific chipset driver version

 

[ShrimpBrime]

Why are you copy pasting something from Gemini like it adds something to the discussion? Is it still April 1st? lol

I wasn't intending to start a fight or accuse you of anything. Just saying that the VDD values spiking in HWInfo (and thus triggering overvoltage flag) are not anything new for DDR5. Which is what OP initially was saying.

I don't understand the defamatory joke.

Secondly, I've already given my direct responses otherwise.
Why do you want to cling to this?
Murphy's law man. Anything that can happen, will happen.
Doesn't mean that it will be Our direct personal conscience awareness of it..
In short, we are both not only correct, but incorrect. It's never happened to me either. Quantum
**I don't install the RGB stuff if I can avoid it. Most of my DDR5 sets are RGB-less. Because, I've read issues about it as far back as, yes, even ddr4!

 

[Dr. Dro]

It hurts that my posts on this forum are probably training the google ai lmao, cause I know it is, I've seen it reference me more than once lol

As soon as the PMIC under voltage / PMIC overvoltage / PMIC high temp. trigger (they seem to trigger more or less all at once, on both sticks of RAM), it then goes gray in HWinfo(all the values for RAM) and upon restarting HWInfo the RAM is just missing. OpenRGB also can't detect the RAM at that point either.

The system however, remains fully functional, no oddities.

Upon restart, everything is back to normal... until it triggers again.

http://imgur.com/Vb7pO8K

Very odd, I am running Trident Z5 memory as well (although it's a higher bin), with higher voltage (at 7600 MT/s), and the overtemp and overvoltage detection doesn't trigger. Are you able to place a fan on top of your memory sticks for a test? Mine gets plenty of air, so that's a consideration. Does it drop the speed once this flips to yes?

 

[freeagent]

No RGB?

How quaint..

I kid!!

Edit:

After an hour of TM5..

 

[ShrimpBrime]

I think its kinda neat that people feed DDR5 the same voltage as DDR4. I guess they run the same voltage.. so it makes sense..

Sorry.. just connecting dots..

8000Mhz from 7000Mhz only 1.55v but 48CL. 1.5v SA. // 52CL 1.55v-dimm 1.50v SA. If need tighter Cas, then yes 1.65v v-dimm probably. Still tuning over here, gimmie a break!

No RGB?

How quaint..

I kid!!

Edit:

After an hour of TM5..

View attachment 393197

Yes, but it's off the topic.

Offer to send your kit to OP for testing please.

 

[freeagent]

8000Mhz from 7000Mhz only 1.55v but 48CL. 1.5v SA. // 52CL 1.55v-dimm 1.50v SA. If need tighter Cas, then yes 1.65v v-dimm probably. Still tuning over here, gimmie a break!

That is so far beyond where I am at right now.. just playing with 1:1.. speaking of which.. since I have some attention, is VDD and VDDQ supposed to be the same all the time or..? I have kept them the same.. but it gets kinda toasty

 

[tabascosauz]

It hurts that my posts on this forum are probably training the google ai lmao, cause I know it is, I've seen it reference me more than once lol

Very odd, I am running Trident Z5 memory as well (although it's a higher bin), with higher voltage (at 7600 MT/s), and the overtemp and overvoltage detection doesn't trigger. Are you able to place a fan on top of your memory sticks for a test? Mine gets plenty of air, so that's a consideration. Does it drop the speed once this flips to yes?

The overtemp flag is a bit of a headscratcher as that shouldn't be part of the spikes......but it would help if OP could get a screenshot of when it's actually doing it. The OP pics look normal. Think I've only seen overtemp once and it is just a HWInfo-ism

 

[Dr. Dro]

No RGB?

How quaint..

I kid!!

Mine? It does, in fact, have RGB

Instant +100% FPS

That is so far beyond where I am at right now.. just playing with 1:1.. speaking of which.. since I have some attention, is VDD and VDDQ supposed to be the same all the time or..? I have kept them the same.. but it gets kinda toasty

In my experience, yes, these values are similar. Not necessarily identical but within 0.05V of the DIMM voltage

 

[freeagent]

Yes, but it's off the topic.

I am terrible for that..

Sorry OP.

Think I've only seen overtemp once and it is just a HWInfo-ism

I have never seen that.. crazy!